448 Annals ofthe Rheumatic Diseases 1992; 51: 448-455
Effects of tiaprofenic acid (Surgam) on cartilage
proteoglycans in the rabbit joint immobilisation Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from model
Isabelle Meyer-Carrive, Peter Ghosh
Abstract pressive deformation.4 5 Biochemical and histo- A well established model of arthritis induced chemical studies of cartilage from osteoarthritic in rabbit knee joints by immobilisation in full joints have shown that proteoglycans are lost extension for 30 days was used to evaluate the from the matrix in the more advanced stages of in vivo effects of 2-5, 5-0, and 10-0 mg/kg body the disease."8 The depletion of these macro- weight oftiaprofenic acid on articular cartilage molecules results in a decline in the resilience of proteoglycans. The drug was given sub- the tissue, and the mechanical stresses trans- cutaneously every 24 hours during the entire mitted across the joint may contribute to the immobilisation period. Immobilised animals progression of pathological changes in cartilage not treated with drugs and normal animals and subchondral bone. were used as controls. In the non-drug treated Non-steroidal anti-inflammatory drugs immobilised animals articular cartilage showed (NSAIDs) are extensively used in the manage- evidence of surface damage accompanied ment of osteoarthritis, but there is continuing by synovial hypertrophy and effusion. Proteo- debate about their beneficial or adverse effects glycan concentrations were reduced in on cartilage integrity."' Some laboratory and cartilages of these joints and the incorporation clinical evidence9 11-14 suggests that several of of 35SO24 into macromolecular proteoglycans the older NSAIDs may suppress the biosynthesis was higher than in cartilages of non-im- of proteoglycans in cartilage, thereby accel- mobilised controls. Gel filtration chroma- erating its failure rate. On the other hand, other tographic studies of both resident and 35S data indicate that the effect of certain NSAIDs labelled proteoglycans isolated from im- on cartilage is innocuous at therapeutically used mobilised joint cartilage showed reduced concentrations,'5 16 or that these drugs may aggregation and the presence of degraded even provide a protective role by support- proteoglycan subunit species. Whereas the ing macromolecule biosynthesis within joint administration of 10,0 mg/kg tiaprofenic acid tissues.'5 17 These latter studies have generally every 24 hours to immobilised animals focused on diclofenac, piroxicam, and tiapro- http://ard.bmj.com/ exacerbated the degradation and loss of fenic acid and, for the most part, have been proteoglycans from joint cartilages, 5-0 mg/kg confined to in vitro investigation. tiaprofenic acid provided some protection of In this study we considered the in vivo these macromolecules, as shown by the con- effects of various doses of tiaprofenic acid on centrations and extractability ofproteoglycans cartilage proteoglycans in a rabbit experimental from cartilages, which were comparable with osteoarthritis. The results obtained indicated,
those from non-immobilised controls. A in this model at least, that the loss or preservation on September 27, 2021 by guest. Protected copyright. high incorporation of 35S into proteoglycans of proteoglycans in articular cartilage is was demonstrated, together with reduced dependent on the amount of drug given. catabolism of proteoglycans, indicating pre- servation of chondrocyte anabolic activity. At a tiaprofenic acid dose of 2.5 mg/kg, however, Materials and methods no beneficial effects on cartilage proteoglycans COMPOUNDS could be shown. Carrier free H235SO4 (specific activity 1-6 TBq/mg) was obtained from Du Pont Medical Products (Wilmington, DE, USA). Guanidine hydrochloride (grade 1), trizma base (reagent Although the cause of osteoarthritis is presently grade (TRIS)), bovine serum albumin, phenyl- unknown, there is general agreement`3 that methylsulphonyl fluoride, benzamidine hydro- failure of articular cartilage represents an early chloride, L-hydroxyproline, L-cysteine free and critical step in the progression of this base, D-glucuronolactone and hyaluronic acid disorder. Articular consists (grade 1, from human umbilical cord), whale Raymond Purves cartilage essentially Research Laboratories, of chondrocytes embedded in a hydrated gel of and shark chondroitin-4-sulphate and chon- (University of Sydney), proteoglycans which are constrained by an droitin-6-sulphate, Coomassie brilliant blue G, Royal North Shore integrated network of collagen fibrils. The and calf thymus DNA were purchased from Hospital of Sydney, St Leonards, NSW 2065, macromolecular organisation of the cartilage Sigma Chemical (St Louis, MO, USA). N- Australia extracellular matrix is complex and is still Ethylmaleimide was obtained from Calbiochem- I Meyer-Carrive the subject of investigation, but it has been Behring Australia (Sydney, Australia), papain P Ghosh established that the entrapment of the hydro- (papainase [EC 3.4.22.2] (puriss) from Fluka Correspondence to: philic proteoglycans within the collagen AG (Buchs, Switzerland), and Sepharose CL- Professor Ghosh. 2B was Pharmacia Accepted for publication meshwork confers to cartilage its hydroelastic supplied by (South Seas) 18 June 1991 properties and ability to recover from com- (Sydney, Australia). All other chemicals and Effects oftiaprofenic acid on cartilage proteoglycans 449
reagents were analytical grade or the highest Group 6: the immobilised joints of six rabbits grade available, unless specified. that received 5 mg/kg tiaprofenic acid sub-
cutaneously every 24 hours during the im- Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from mobilisation period (5 0 mg treated group). PROCEDURES Group 7: the contralateral joints of the six Induction of cartilage degeneration in the rabbit group 6 animals. knee joint by immobilisation Group 8: the immobilised joints of three The experimental series comprised 31 mature rabbits given 2-5 mg/kg tiaprofenic acid sub- (9-12 months old) New Zealand white rabbits. cutaneously during the immobilisation period Rabbit joints were immobilised by a modifica- (2-5 mg treated group). tion of published techniques.8 19 Briefly, two Group 9: the contralateral joints of the three plastic concave splints (220x25x2 mm; 200x group 8 animals. 20x2 mm) were used. The larger splint was applied to the ventral aspect of the left leg of the rabbits from the thigh to the distal end of the DISSECTION AND ANALYSIS OF TISSUES limb, which was maintained in full extension. Rabbit joints were opened and the articular The splint was attached to the limb by wrapping cartilage was removed from femoral condyles, five turns of leucoplast bandage (Johnson and suprapatellar groove, and tibial plateaux with a Johnson, Sydney, Australia) around the splint surgical blade. The cartilage from each of these and leg. The second splint was similarly fixed regions was pooled, finely diced, transferred to but to the dorsal aspect of the limb. The tension preweighed stoppered tubes, and lyophilised. in the bandage did not impair the blood supply Aliquots of the pooled anhydrous cartilage were to the leg. This procedure allowed ± 5° of analysed for hexuronic acid content, after movement of the joint within the splint. The papain digestion20 by the method of Blumen- rabbits were allowed to move freely during the krantz and Asboe-Hansen,21 using glucurono- immobilisation period, checked daily, and the lactone as a standard. The sulphated glyco- bandage replaced if damaged. saminoglycans of column fractions were measured by the methods of Farndale et a122 Drug administration and isotopic labelling of with chondroitin-6-sulphate as a standard. The articular cartilage proteoglycans radioactivity of proteoglycan samples, or The rabbits were treated prophylactically with column fractions, was determined by liquid 10-0, 5-0, or 2-5 mg/kg tiaprofenic acid during scintillation spectrometry. All analyses were the entire immobilisation period. The drug as conducted in triplicate. the trometamol salt injectable form (Roussel UCLAF, Sydney, Australia) was given sub- cutaneously every 24 hours over the 30 day EXTRACTION AND PURIFICATION OF immobilisation period, so that each animal PROTEOGLYCANS received a total of 29 injections. Twenty four Proteoglycans were dissociatively extracted http://ard.bmj.com/ hours before they were killed the drug treated from aliquots of diced wet articular cartilage immobilised group, the immobilised group not with 10 volumes of freshly prepared buffered treated with drugs, and the non-immobilised 4 M guanidine hydrochloride containing the control group not treated with drugs were protease inhibitors 25 mM EDTA, 10 mM N- injected intravenously with H235SO4 in physio- ethylmaleimide, and 1 mM benzamidine HC1 in logical saline (1-0 ml) (92-5 MBq/kg) through 50 mM TRIS at pH 7*4 for 48 hours at 4°C, as the marginal ear vein. The animals were killed described by Oegema etal.23 After centrifugation on September 27, 2021 by guest. Protected copyright. by an intravenous injection of sodium pento- (20 minutes at 2000 g) the residues were washed barbitone (75 mg/kg) into the marginal ear vein. twice with a small volume (-0 5 ml) ofextraction buffer, and the supernatants and washings were pooled. The residues were washed thoroughly EXPERIMENTAL GROUPS EXAMINED (48 hours extraction) with distilled water and The following groups were examined: lyophilised before papain digestion.20 The Group 1: nine rabbits that were not im- residues were analysed for hexuronic acid mobilised and did not receive any drug (external content,2 as described above. The percentage control group). of proteoglycans extracted from the cartilage Group 2: nine rabbits with immobilised joints was calculated as the hexuronic acid content in that received no drug during the experiment the extract x 100/total hexuronic acid concen- (non-drug treated immobilised joint group). trations in residues and extracts. Group 3: the contralateral joints of the nine The 4 M guanidine hydrochloride extracts group 2 animals. Our previous experiments'9 were dialysed in the presence of hyaluronic acid using this model had shown that the cartilage of (2% (wt/wt) of extract hexuronate content) contralateral joints was also influenced by the overnight at 4°C against nine volumes 0-5 M immobilisation procedure/(non-drug treated sodium acetate (pH 6-4) containing the same contralateral joint group). enzyme inhibitors as used for the extraction Group 4: the immobilised joints of four buffer,23 thereby reducing the guanidine hydro- rabbits that were given 10-0 mg/kg tiaprofenic chloride concentration to 0-4 mol/1.24 Solid CsCl acid subcutaneously every 24 hours throughout was then added to produce a starting density of the immobilisation period (10-0 mg treated 1 6 g/m124 and the solution subjected to ultra- group). centrifugation for 24 hours at 6°C at 50 000 rpm Group 5: the contralateral joints of the four with a Sorval OTD65B centrifuge and TV-865B group 4 animals. vertical rotor, as described previously.25 The 450 Meyer-Carrive, Ghosh
centrifuge tubes were manually fractionated of articular cartilage of the non-drug treated into six 700 1t fractions with an automatic immobilised joints (group 2) was significantly
pipette, starting at the top of the tube. Aliquots lower than that of the contralateral joints Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from of these fractions were analysed for protein by (group 3) (p-<00005) and that of the non-im- the bicinchoninic acid method26 and sulphated mobilised controls (group 1) (p-0O0005). The glycosaminoglycans by the method of Farndale hexuronate content of the cartilages of the et al.22 The density of the fractions was deter- contralateral free joints of the immobilised mined by weighing an aliquot of known volume. animals (group 3) was also significantly lower Ninety per cent of the proteoglycans were found (p-<0005) than that of the non-immobilised in the bottom two fractions (density - 1 65). control (group 1) animals. These were therefore combined and called the The loss of proteoglycans from the knee joint Al fractions, according to the terminology of articular cartilage of immobilised joints was Heinegard.27 enhanced when animals were given 10 mg/kg tiaprofenic acid (group 4). This loss was signi- ficantly greater than for the group 2 joints SEPHAROSE CL-2B GEL CHROMATOGRAPHY (p-0O0005). Similarly, the contralateral joint Aliquots of the Al fractions, which contained cartilage values (group 5) were significantly 40-100 [tg of hexuronic acid, were chromato- lower than those for the non-drug treated graphed on a Sepharose CL-2B column (20x 1 control group (group 3) (p-0*0005). cm), eluted with 0 5 M sodium acetate (pH In contrast with the animals who received 10 6 8). Flow through the column was 10 ml/h, mg/kg tiaprofenic acid, the hexuronate concen- and the column was calibrated with highly trations in cartilages of animals receiving 5*0 polymerised DNA (V0), H235S04 (V,), and a mg/kg tiaprofenic acid (groups 6 and 7) were papain digested cartilage sample (chondroitin higher than those found in cartilages of the non- sulphate peptides). Fractions of 250 ,ul were drug treated immobilised groups (groups 2 and collected, 100 p1 aliquots counted for radio- 3). These differences were significant at the activity (24 hour synthesised proteoglycans), p-0*005 level (fig 1). On the other hand, and 10-100 p1 aliquots were used to determine the hexuronate concentrations in cartilages of total sulphated glycosaminoglycans, using the immobilised animals treated with 2-5 mg/kg method of Farndale et al.22 tiaprofenic acid (groups 8 and 9) were indistin- guishable statistically from values in non-drug STATISTICAL ANALYSIS OF RESULTS treated controls (groups 2 and 3). The data obtained were analysed either by a paired Student's t test between groups of the same animals (immobilised and contralateral) or PROTEOGLYCAN EXTRACTABILITY an unpaired Student's t test between groups The proportion of proteoglycans extracted from of different animals. Differences between cartilages of non-drug treated immobilised groups were considered to be significant when and contralateral joints (groups 2 and 3) was http://ard.bmj.com/ p-O O5. significantly lower than that of the corres- ponding non-immobilised controls (group 1) (p-00005; p-0 005 respectively) (fig 2). Results Apart from cartilages of animals receiving PROTEOGLYCAN CONTENT (AS HEXURONATE) 5-0 mg/kg tiaprofenic acid, the proportion of Figure 1 shows that the hexuronic acid content proteoglycans extracted from groups 4, 8, 5,
and 9 was similar to that from the non-drug on September 27, 2021 by guest. Protected copyright. treated immobilised groups (groups 2 and 3) (fig 2).
d -ol
j~~ a
X_-nlt r L) laterali I' "..1 Figure I Mean (SEM) hexuronic acid concentration (Mglmg dry weight) ofrabbit kneejoint articular cartilage ''m from normal (GI); immobilised (G2): contralateral to (D immobilised Ln (G3); immobilised and drug treated 0 (,1: 10 mglkg) (G4); contralateraljoint to G4 (GS); -.. '' '. immobilised and drug treated (a: 5 mglkg) (G6); Contral n:olhHiis- ..ot . contralateraljoint to G6 (G7); immobilised and drug treated (M: 2-5 mglkg) (G8); contralateraljoint to G8 (G9) animals. Significant differencesfound between groups were: GI >G2 Figure 2 Mean (SEM) percentage extractability of (p-00005); GI>G3 (p<0-005); proteoglycans (4 0M guanidine hydrochloridefrom rabbit GJ>G4 (p-0-0005); GI>GS (p-0-0005); kneejoint cartilages. Group coding asforfig 1. Significant Gl>G6 (ps0-01); GI>G7 (p-0-025); differencesfound between groups were: GI>G2 Gl>G8 (p-0-0005); GJ>G9 (ps0-05); G2 35S LABEL.LED PROTEOGLYCAN SPECIFIC ACTIVITY SEPHAROSE CL-2B CHROMATOGRAPHY AND Figure 3 shows the specific activity of 35S HYALURONIC ACID BINDING STUDIES labelled proteoglycans isolated from joint Figures 4 and S show the chromatographic Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from cartilages of normal and experimental animals. profiles of the resident proteoglycans as deter- Joint imnnobilisation increased the incorporation mined by the assay of Farndale et al,22 and of 35S irnto the extractable proteoglycans by newly synthesised proteoglycans, as measured about 50'% of the control values (group 1 v by 35S radioactivity, from the cartilage of group 2) (p<0005). Treatment with 10 mg/kg normal and experimental animals. Whereas the tiaprofenic acid decreased the incorporation of polydispersity and extent of aggregation of the 35S into p.roteoglycans to within non-immobilised newly synthesised (35S labelled proteoglycans) control 14evels (group 1), but animals given S and resident populations of proteoglycans were mg/kg off the drug (groups 6 and 7) showed similar in normal cartilage (figs 4A and SA), enhanced biosynthetic activity in both im- the proteoglycans isolated from the cartilage mobilisecI and contralateral joints, as shown by of immobilised joints (including contralateral the highoer specific activity value of isolated joints), showed decreased aggregation for both proteoglyrcans (fig 3). resident and newly synthesised species (figs 4B and SB). The aggregation was significantly lower than in group 1 (p.0 0005) (table). Moreover, the hydrodynamic range of the proteoglycans isolated from the immobilised cartilage was markedly altered, a major popu- X 1600 lation eluting with Ka, values similar to those of the chondroitin sulphate peptides (Kav,0-85) ° 1400- (figs 4B and SB, table). u x m X Proteoglycans isolated from the cartilage of = 1200 - ~~~~~animals given 10 mg/kg tiaprofenic acid showed 0 @ CD1000 similar aggregation characteristics to those of the proteoglycans of the immobilised but non- L 800- drug treated controls. The non-aggregatable proteoglycan population, however, was of Conltrol Immllobilised Immlllobilisec overall smaller hydrodynamic size (figs 4C and conitralateral 5C). On the other hand, the proteoglycans Figure 3 Mean (SEM) specific activity (as dpml4g extracted from the cartilage of immobilised hexuronate of3S labelledproteoglycans isolatedfrom animals given 5 mg/kg tiaprofenic acid showed rabbit kneeJoint articular cartilages. Group coding asfor aggregation characteristics which were com- fig 1. Signi,ificant differencesfound between groups were: GI A 0R %1%2CS on September 27, 2021 by guest. Protected copyright. 024 0-22 E E c c 0 0 In 0-20- Cis c] 0180-18 0 0 U-i ll D I . * I 9. * . I U-1 IX . . . -0-4 0-0 0-4 0-8 1-2 -0-4 0-0 0-4 0-8 1-2 0 E E c C 0 0 ,I* 0 0 -0-4 0-0 0-4 0-8 1-2 -0-4 0-0 0-4 0-8 1-2 Figure 4 Sepharose CL-2B chromatograms ofresident proteoglycans (as Farmdale et a122 positive material) isolatedfrom rabbit kneejoint cartilages of(A) non-immobilised non-drug treated animals; (B) immobilised (-E*) and contralateral (O -O) ofnon-drug treated animals; (C) immobilised (m- ) and contralateral (O-O) ofanimals treated with 10-0 mglkg oftiaprofenic acid; and (D) immobilised (E-E) and contralateral (O-O) ofanimals treated with S 0 mglkg tiaprofenic acid. CS represent the elution position ofchondroitin sulphate peptides (K,,,,,=0-85). 452 Meyer-Carrive, Ghosh 400 ui Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from LoE E 0. 'a -0-4 0.0 0-4 0-8 1-2 -04 0-0 0-4 0-8 1-2 Ln Ln - E 0. -oE 'a a 100 -0-4 0-0 0-4 0-8 1-2 -0-4 0.0 0-4 0-8 1-2 Kav Kav Figure 5 Sepharose CL-2B chromatogram of35" labelled proteoglycans isolatedfrom rabbit kneejoint cartilages. Panels and symbols identical to those used in fig 4. was much less than the corresponding fractions Discussion of the non-drug treated immobilised group (figs Using a modification of the immobilisation 4 and 5, table). With animals given 2 5 mg/kg technique of Langenskiold et al, 8 we induced tiaprofenic acid no significant difference was reproducible changes in the articular cartilage found relative to the non-drug treated im- and periarticular structures of rabbit joints after mobilised groups (data not shown). four weeks. The morphological changes seen Aggregation (as percentage of total) of resident and newly synthesised proteoglycans isolated from rabbit joint articular cartilages Regimen Group 3SS labelled proteoglycans (dpm) Resident proteoglycans http://ard.bmj.com/ Mean (SEM) Significant difference Mean (SEM) Significant difference between groups between groups Non-immobilised control (n=9) 1 40 4 (0 63) 41 7 (0-9) Non-drug treated immobilised 2 31-3 (2 0) G1>G2 27 5 (1 8) G1>G2 joint (n=9) (p Non-drug treated contralateral 3 23-9 (2 8) G1>G3 36-0 (1 5) G1>G3 on September 27, 2021 by guest. Protected copyright. joint (n=9) (pO 0005) (pO.005) G3 *Farndale et al.22 tNSD=not statistically significant from. Effects oftiaprofenic acid on cartilage proteoglycans 453 were decreased lustre of the cartilage surface acid given to rabbits in this model was reduced with some pitting, subchondral hyperaemia, to 5-0 mg/kg the proteoglycans in the cartilage thickening of the synovium, and increased of immobilised joints were protected. Further- Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from synovial effusions (data not shown). Similar more, from the 35S specific activity and chroma- changes have been described by others.'8 28-32 tography studies it was clear that suppression of It has been reported that proliferation of proteoglycan biosynthesis by the drug at this synovial lining cells and infiltration of mono- dose did not occur. In fact, the incorporation of nuclear cells occurs within a week of im- 35S into the newly synthesised population was mobilisation of rabbit joints.32 This is then higher than in the non-drug treated immobilised followed by cartilage surface changes, softening, controls (fig 3). At a dose of 2-5 mg/kg and chondrocyteproliferation at the joint margins tiaprofenic acid failed to influence positively the accompanied by subchondral changes.'8 28-32 loss of proteoglycans from cartilage of immo- We found that the cartilage ofthe immobilised bilised joints. Although we had chosen a joints contained a reduced amount of proteo- prophylactic drug protocol, it is likely, on the glycans, which were less extractable. From the basis of other investigations using osteoarthritic gel filtration studies, this seems to be due to cartilage38 39 and a sheep model of osteo- their enhanced catabolism both at the hyalu- arthritis,40 that tiaprofenic acid is also effective ronate binding region and the chondroitin once cartilage damage has started. sulphate-rich regions of the proteoglycan core The only other drugs reported to show a protein. This was evident from the reduced protective effect on articular cartilage in this ability of proteoglycans to form aggregates in animal model were pentosan polysulphate the presence of hyaluronic acid (table) and the (Cartrophen)'9 and glycosaminoglycan poly- inclusion of degraded non-aggregating proteo- sulphate (Arteparon):4' These agents are not glycan products into the Sepharose CL-2B gel NSAIDs, however, but oversulphated poly- (figs 4 and 5). It was noteworthy that this saccharides and were effective at 10 mg/kg but increased breakdown of proteoglycans occurred not 5 mg/kg when given intramuscularly over in cartilages of both immobilised and contra- the four week immobilisation period.'9 41 It lateral joints; this was particularly true for the should be noted that both these polysulphated newly synthesised 35S labelled proteoglycans polysaccharides have shown useful antiarthritic (fig 5). We assume that this situation arises from activities when used in other animal models as the overall impairment of movement incurred well as clinically.42 4 by the immobilised limb, which introduces Tiaprofenic acid is a potent inhibitor of abnormal mechanical stresses across the contra- prostaglandin synthetase, being twice as active lateral joint of the animal. The finding of as indomethacin or diclofenac in the bovine increased incorporation of 35S into proteo- seminal vesicle assay.44 Measurement of prosta- glycans of cartilage of immobilised rabbit joints glandin E2 concentrations and phospholipase was consistent with the report of Videman et A2 in synovial fluids of patients with rheuma- 01.33 Although this indicates an increased toid arthritis before and after oral treatment http://ard.bmj.com/ metabolic response by chondrocytes to the with tiaprofenic acid has confirmed that depletion ofproteoglycans from the extracellular prostaglandin activity can be suppressed by this matrix, similar changes do not occur in the drug in human tissues.45 46 Furthermore, the cartilage of immobilised sheep34 or in canine catabolism of human osteoarthritic cartilage joints,35 suggesting that the reaction of rabbit maintained in culture was shown to be reduced cartilage to immobilisation may be more severe when tiaprofenic acid at 2-6 [tg/ml was incor- than in other species. porated into the medium.38 39 As this effect on September 27, 2021 by guest. Protected copyright. Administration of 10 0 mg/kg tiaprofenic acid could be ameliorated by addition of prosta- to the rabbits over the entire immobilisation glandin E2 or dibutyl cyclic AMP to the drug period resulted in a marked depression in treated cultures it was suggested39 that tia- proteoglycan biosynthesis to joint cartilages profenic acid suppresses prostaglandin E2 and (fig 3), accompanied by a loss of these macro- cyclic AMP synthesis by chondrocytes. These molecules from the matrix (figs 4 and 5). The studies led us to suggest that administration hydrodynamic size and polydispersity of the of 5 mg/kg tiaprofenic acid in the rabbit osteo- proteoglycans isolated were again consistent arthritis model might achieve tissue concen- with proteolytic cleavage of proteoglycan core trations able to modulate the biosynthesis of protein at the hyaluronate binding and the prostaglandin E2 and cyclic AMP by both chondroitin sulphate-rich region (figs 4 and 5). synovial cells and chondrocytes. As it is known Diclofenac when given orally at 1-5 mg/kg over that prostaglandin E2 and cyclic AMP stimulate 17 days in the same animal model as used here the production of lymphokines47 and metallo- was also found to exacerbate the loss of proteo- proteinase48 by chondrocytes and synovial cells glycans (measured as glycosaminoglycans) from it would be reasonable to assume that the immobilised joint cartilage.36 In this study, reduced catabolism of cartilage proteoglycans in however, the 35S radiolabelling experiments immobilised joints in the presence of tiaprofenic showed that proteoglycan biosynthesis was not acid proceeds through such pathways. altered by administration of the drug. In another Several independent investigations have study37 indomethacin when given intraperi- shown that arachidonic acid metabolites, tonally at 10 mg/kg to immobilised rabbits for including prostaglandin E2, can regulate inter- 17 days did not increase the loss ofproteoglycans leukin 1,47 49¶50 tumour necrosis factor,5' 52 from immobilised joint cartilages. and superoxide free radical53 activity in inflam- In view of these data it was therefore of some matory cells, and it has been suggested that interest to find that when the dose of tiaprofenic blockade of prostaglandin E2 synthesis by 454 Meyer-Camve, Ghosh NSAIDs may promote, through a feedback 2 Thompson RC Jr, Oegema T R Jr. Metabolic activity of articular cartilage in osteoarthritis. J Bone joint Surg [Am] loop, the production of interleukin 1 by 1979; 61: 407-16. 3 Fassbender H G. Joint destruction in various arthritic chondrocytes.54 In addition, it has recently been Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from disease. In: Kuettner K E, Schleyerback R, Hascall V-C, shown55 that the strong inhibitor of prosta- eds. Articular cartilage biochemistry. New York: Raven glandin E2 biosynthesis, piroxicam, when used Press, 1986: 371. 4 Kempson G E, Muir M, Swanson S A V, et al. Correlations at 5 mg/kg in mice, enhanced lymphocyte between the stiffness and the chemical constituents of proliferation and production of lymphokines in cartilage on the human femoral head. Biochim Biophys Acta 1970; 215: 70-7. mononuclear phagocytes. Taken together, these 5 Maroudas A, Ziu L, Weisman N, Venn M F. Studies of studies strongly suggest that NSAIDs can both hydration and swelling pressure in normal and osteoarthritic cartilages. Biorheology 1985; 22: 87-122. inhibit the production of proinflammatory 6 Mankin H J, Lippiello L. Biochemical and metabolic prostaglandins and stimulate aspects of the abnormalities in articular cartilage from osteo-arthrotic human hips. J Bone Joint Surg [Aml 1970; 52: 424-34. immune response, such as the production of 7 Sweet M B E, Thonar E J-M A, Immelman A R, Solomon L. interleukin 1 and other cytokines. Biochemical changes in progressive osteoarthrosis. Ann Rheum Dis 1977; 36: 387-98. We have formulated a working hypothesis to 8 Poole A R. Changes in the collagen and proteoglycan of explain the results obtained in this study. We articular cartilage in arthritis. Rheumatology 1986; 10: 316-71. suggest that the inhibition of proteoglycan 9 Doherty M. Chondroprotection by non-steroidal anti- synthesis and enhanced catabolism of proteo- inflammatory drugs. Agtn Rheum Dis 1989; 48: 619-21. 10 Ghosh P. Chondroprotective drugs and osteoarthritis. Ann glycans obtained in immobilised rabbit joints Rheum Dis 1990; 49: 338-9. after administration of high dose tiaprofenic 11 Ghosh P, Brooks P. Chondroprotection exploring the concept. J Rheumatol 1991; 18: 161-6. acid (10 mg/kg) arise from the effect of inter- 12 Ghosh P. Anti-rheumatic drugs and cartilage. Baillieres Clin leukin 1 and, possibly, tumour necrosis factor Rheumatol 1988; 2: 309-38. 13 Herman J H, Appel AM, Khosla R C, Hess E V. The in vitro and superoxide radicals on the function of effect of select classes of nonsteroidal antiinflammatory chondrocytes. Such mediators, particularly drugs on normal cartilage metabolism. J Rheumatol 1986; 13: 1014-8. interleukin 1, are known to inhibit proteoglycan 14 Palmoski M J, Brandt K J. Aspirin aggravates the degenera- and collagen biosynthesis by chondrocytes56 58 tion of canine joint cartilage caused by immobilization. Arthritis Rheum 1982; 25: 1333-42. and promote matrix component degradation 15 Brandt K D. Effects of nonsteroidal anti-inflammatory drugs either directly, or through the release of on chondrocyte metabolism in vitro and in vivo. Am3r Med 1987; 83: 29-34. metalloproteinases. It is further proposed that 16 Muir H, Carney S L, Hall L G. Effect of tiaprofenic acid and interleukin 1 and other mediators are increased other NSAIDs on proteoglycan metabolism in articular cartilage explants. Drugs 1988; 35 (suppl 1): 15-23. through the feed back loop51-54 which arises as 17 Collier S, Ghosh P. Evaluation of the effects of antiarthritic a consequence of the strong suppression of drugs on the secretion of proteoglycans by lapine chondro- cytes using a novel assay procedure. Ann Rheum Dis 1989; prostaglandin E synthesis (including anti- 48: 372-81. inflammatory prostaglandin E1) by tiaprofenic 18 Langenskiold A, Michelsson J-E, Videman T. Osteoarthritis of the knee in the rabbit produced by immobilization. acid when used in rabbits at high concentration Attempts to achieve a reproducible model for studies on (10 mg/kg). On the other hand, at the lower pathogenesis and therapy. Acta Orthop Scand 1979; 50: 1-14. dose of 5 0 mg/kg tiaprofenic acid, prosta- 19 Golding J C, Ghosh P. Drugs for osteoarthrosis. I. The glandin and E2 activity in immobilised joints, effects of pentosan polysulphate (SP54) on the degradation El and loss of proteoglycans from articular cartilage in a model while modulated, is not repressed to the extent of osteoarthrosis induced in the rabbit knee joint by http://ard.bmj.com/ that interleukin 1 activity is enhanced or the immobilization. Current Therapeutic Research 1983; 33: 173-84. useful effects of prostaglandin E1 abolished. 20 Pearce R H, Mathieson M H. Quantitative isolation of Hence catabolic processes are controlled but purified acidic glycosaminoglycans from rat skin. Can7 Biochem 1976; 45: 1565-76. chondrocyte anabolic activity is not com- 21 Blumenkrantz N, Asboe-Hansen G. New method for quanti- promised. This proposal is at present highly tative determination of uronic acids. Anal Biochem 1973; 54: 484-89. speculative, but experiments are in progress to 22 Farndale R W, Sayers C A, Barrett A J. A direct spectro- test its validity. photometric microassay for sulfated glycosaminoglycans in on September 27, 2021 by guest. Protected copyright. cartilage cultures. Connect Tissue Res 1982; 19: 247-48. Despite our limited knowledge of the dose 23Oegema TR Jr, Hascall V C, Eisenstein R. Characterization dependent mechanism of action of tiaprofenic of bovine aorta proteoglycan extracted with guanidine hydrochloride in the presence of protease inhibitors. J Biol acid in vivo it is clear from the present studies Chem 1979; 254: 1312-8. that this NSAID, in contrast with aspirin,59 24 Bushell G R, Ghosh P, Taylor T K F, Sutherland J M, Braund K G. The effect of spinal fusion on the collagen and diclofenac,36 and indomethacin,37 can protect proteoglycans of the canine intervertebral disc. J Surg Res articular cartilage in this animal model of 1978; 25: 61-9. 25 Cole T-C, Ghosh P, Taylor T K F. Variations of the arthritis when used at an appropriate dose. In proteoglycans of the canine intervertebral disc with this regard it is noteworthy that our pharmaco- ageing. Biochim Biophys Acta 1986; 880: 209-19. 26 Smith P K, Krohn R I, Hermanson G T, et al. Measurement kinetic studies have shown that tiaprofenic acid of protein using bicinchoninic acid. Anal Biochem 1985; at 5 0 mg/kg given subcutaneously in the rabbit 150: 76-85. 27 Heinegard D. Extraction, fractionation and characterization achieved peak serum concentrations of 25-30 of proteoglycans from bovine tracheal cartilage.' Biochim (Meyer-Carrive I et al, unpublished data) Biophys Acta 1972; 285: 181-92. [tg/ml 28 Videman T, Michelsson J-E, Langenskiold A. The develop- which was comparable with the concentration ment of radiographic changes in experimental osteoarthritis found in man after daily oral administration of provoked by immobilization of the knee in rabbits. IRCS Medical Science 1977; 5: 68. 600 mg of this drug.60 29 Michelsson J E. The effect of ibuprofen on the thickening, stiffening and development of degenerative changes in the rabbit knee following immobilization. Scand J Rheumatol We thank Ms Lynne Campbell for her assistance in preparing 1980; 9:141-4. this manuscript, Dr Arthur Bolt for valuable discussions on the 30 Videman T. Experimental osteoarthritis in the rabbit- animal experiments, Ms Fiona Brook for her help with animal comparison of differentperiods of repeated immobilization. handling, and Ms Anne Darvodelsky in assisting with the Acta Orthop Scand 1982; 53: 339-47. proteoglycan aggregation studies. We also acknowledge with thanks financial support for this project, in the form of a 31 Candolin T, Videman T. Surface changes in the articular postgraduate scholarship to Isabelle Meyer-Carrive, from cartilage of rabbit knee during immobilization. A scanning Roussel UCLAF Australia Pty Limited. electron microscopic study of experimental osteoarthritis. Acta Pathol Microbiol Immunol Scand [Al 1980; 88: 291-7. 32 Michelsson J E, Hunneyball I M. Inflammatory involvement McDevitt C A, Gilbertson E M, Muir H. An experimental in rabbit knee followingimmobilization and resulting in model of osteoarthritis: early morphological and bio- osteoarthritis. Scand J Rheumatol 1984; 13: 273-81. chemical changes. Bone Joint Surg[Br] 1977; 59: 24-35. 33 Videman T, EronenI, Friman C. Glycosaminoglycan Effects oftiaprofenic acid on cartilage proteoglycans 455 metabolism in experimental osteoarthritis caused by 47 Lewis G P. Immunoregulatory activity of metabolites of immobilization: the effects of different periods of im- arachidonic acid and their role in inflammation. Br Med mobilization and follow-up. Acta Orthop Scand 1981; 52: Bull 1983; 39: 243-8. 11-21. 48 Malemud C, Sokoloff L. The effect of prostaglandins on Ann Rheum Dis: first published as 10.1136/ard.51.4.448 on 1 April 1992. Downloaded from 34 Caterson B, Lowther D A. Changes in the metabolism of the cultured lapine articular cartilage. Prostaglandins 1977; 13: proteoglycans from sheep articular cartilage in response to 845-59. mechanical stress. Biochim Biophys Acta 1978; 540: 412-22. 49 Kunkel S L, Chensue S W, Phan S H. Prostaglandins as 35 Palmoski M, Perricone E, Brandt K D. Development and endogenous mediators of interleukin 1 production. reversal of a proteoglycan aggregation defect in normal .7 Immunol 1986; 136: 186-92. canine knee cartilage after immobilization. Arthnrtis Rheum 50 Knudsen P J, Dinarello C A, Strom T B. Prostaglandins post 1979; 22: 508-17. transcriptionally inhibit monocyte expression of interleukin 36 Eronen I, Videman T. Effects of sodium diclofenac on 1 activity by increasing cyclic adenosine monophosphate. glycosaminoglycan metabolism in experimental osteo- Immunol 1986; 137: 3189-94. arthritis in rabbits. Scand Rheumatol 1985; 14: 37-42. 51 Kunkel S L, Spengler M, May M A, Spengler R, Larrick J, 37 Videman T, Eronen I, Friman C, Ahtiala K, Mattila M J. Remick D. Prostaglandin E2 regulates macrophage-derived Effects of indomethacin on glycosaminoglycan metabolism tumour necrosis factor gene expression. Biol Chem 1988; in the development of experimental osteoarthritis in 263: 5380-4. rabbits. Biochem Pharmacol 1981; 30: 2135-9. 52 Hart P H, Whitty G A, Piccoli D S, Hamilton J A. Control by 38 Pelletier J-P, Cloutier J M, Martel-Pelletier J. In vitro effects IFN-y and PGE2 or TNFa and IL-I production by human of tiaprofenic acid, sodium salicylate and hydrocortisone monocytes. Immunology 1989; 66: 376-83. on the proteoglycan metabolism of human osteoarthritic 53 Sedgwick J B, Berube M L, Zurier R B. Stimulus-dependent cartilage. Rheumatol 1989; 16: 646-55. inhibition of superoxide generation by prostaglandins. Clin 39 Martel-Pelletier J, Pelletier J P. Molecular basis for the action Immunol Immunopathol 1985; 34: 205-15. of tiaprofenic acid on human osteoarthritic cartilage 54 Dingle J T. Cartilage maintenance in osteoarthritis-inter- degradation. Semin Arthritis Rheum 1989; 18 (suppl 1): actions of cytokines. NSAID and prostaglandins in 19-26. articular-cartilage damage and repair. Rheumatol 1991; 18 40 Ghosh P, Holbert C, Read R, Armstrong S, Wilson D. (suppl 28): 30-7. Chondroprotection and NSAIDs. Lancet 1991; 337: 1612-3. 55 Hayns D R, Wright P F, Whitehouse M W, Vernon-Roberts 41 Ghosh P, Golding J. Drugs for osteoarthrosis. II. The effects B. The cyclo-oxygenase inhibitor piroxicam, enhances of Arteparon on proteoglycan aggregation and loss from cytokine-induced lymphocyte proliferation in vitro and in articular cartilage in a model arthropathy in the rabbit. vivo. Immunol Cell Bwil 1990; 68: 225-30. Current Therapeutic Research 1983; 34: 67-80. 56 Tyler J A, Saklatvala J. Pig interleukin 1 (catabolin) induces 42 Burkhardt D, Ghosh P. Laboratory evaluation of anti- resorption of cartilage proteoglycan and prevents synthesis arthritic drugs as potential chondroprotective agents. Semin of proteoglycan and collagen. Br Rheumatol 1985; 24 Arthritis Rheum 1987; 17 (suppl 1): 3-34. (suppl 1): 150-5. 43 Rejholec V. Long term studies of antiosteoarthritic drugs: an 57 Saklatvala J, Sarsfield S J. How do interleukin 1 and tumour asessment. Semin Arthritis Rheum 1987; 17 (suppl 1): necrosis factor induce degradation of proteoglycan in 35-53. cartilage? In: Glauert A M, ed. The control oftissue damage. 44 Deraedt R, Benzoni J, Delvallee F. Pharmacological profile Amsterdam, NY, Oxford: Elsevier, 1988: 97-108. of tiaprofenic acid. Rheumatology 1982; 7: 78-87. 58 Bates E J, Lowther D A, Handley C J. Oxygen free radicals 45 Daymond T J, Rowell F J. Reduction of prostaglandin-E2 mediate an inhibition of proteoglycan synthesis in cultured concentrations in synovial-fluid of patients suffering from articular cartilage. Ann Rheum Dis 1984; 43: 462-9. rheumatoid-arthritis following tiaprofenic acid or indo- 59 Michelsson J-E. Effect of acetyl salicylic acid on thickening of methacin treatment. Drugs 1988; 35 (suppl 1): 4-8. the rabbit knee following immobilization. IRCS Medical 46 Vignon E, Mathieu P, Couprie N, et al. Effects of tiaprofenic Science 1979; 7: 37. acid on interleukin-1, phospholipase A2 activity, prosta- 60 Daymond T J, Herbert R. Simultaneous bioavailability of glandins, neutral protease, and collagenase activity in tiaprofenic acid (Surgam) in serum and synovial fluid in rheumatoid synovial fluid. Semin Arthritis Rheum 1989; 18 patients with rheumatoid arthritis. Rheumatology 1982; 1: (suppl 1): 11-5. 188-93. http://ard.bmj.com/ on September 27, 2021 by guest. Protected copyright.